Foot-and-mouth disease (FMD) is the most economically significant disease of domestic livestock. Cloven-hoofed animals of the order Artiodactyla, such as cattle, pigs, sheep and goats, are susceptible. The causative agent, foot-and-mouth disease virus (FMDV), is an aphthovirus of the Picornaviridae family having a positive sense RNA genome encoding four capsid proteins (VP1-VP4) and non-structural polypeptides. The FMDV genome, the coding region for capsid and non-structural proteins and the processing pathway by which proteins are cleaved is shown schematically in FIG. 1 (taken from Tesar et al., Virus Genes, 1989; 3:29-44).
FMD is controlled by quarantine and destruction of infected and exposed animals, and by vaccination with chemically inactivated virus compositions. A significant difficulty for formulation of these vaccines is the remarkable antigenic diversity of FMDV. Seven distinct serotypes have been described: A, O, C, Asia, and the South African types SAT-1, 2, and 3, each of which can be subdivided into multiple subtypes. Viruses of serotypes A, O, and Asia are most common. Serotype A viruses are the most variable, having more than 30 subtypes.
There is no cross-protection between serotypes, therefore animals recovered from infection with or vaccinated against a virus of one serotype are still susceptible to infection with viruses from the remaining six serotypes. Moreover, the degree of antigenic variation within a serotype is such that a vaccine effective against one subtype may not be protective against another subtype within that same serotype. Accordingly, current vaccines are based on a mixture of inactivated viruses comprising a reference strain of each relevant serotype or subtype, as determined by monitoring the virus strains in local circulation. Periodic surveillance on the antigenic relationship between field isolates, including emerging variants, and the viruses being used for vaccine preparation are required to prevent outbreaks.
Therefore, the production of current vaccines requires that several virus strains must be grown and inactivated, with the potency of each assured, and the product must be reformulated periodically with current field strains to prevent loss of protective efficacy by emergence of new variants. The process of preparing multiple virus strains for maintenance of immunogenic efficacy in the face of antigenic variation results in burdensome problems with strain to strain variability in potency and yield during manufacture.
The need for an inactivated virus vaccine for periodic revision in turn generates a new obstacle to vaccine manufacture. When a vaccine is revised to meet the challenge of an outbreak by a newly evolved field strain, the outbreak strain must be artificially adapted to efficient growth in tissue culture. Furthermore, the selection process for adaptation if newly emerging subtypes to tissue culture can result in diminished protective immunogenicity, resulting in further delays in production. (Barteling and Vreeswijk, Vaccine, 1991; 9: 75-88 Brown, Vaccine, 1992; 10: 1022-1026).
Other significant shortcomings of inactivated FMDV vaccines are associated with issues of biosecurity and stability. The virus must be produced in a high-containment facility to prevent contamination of the immediate environment. This potential hazard limits qualified vaccine suppliers to a few, limits the available supply of vaccine, and hampers efforts to respond to outbreaks in previously FMDV-free regions of the world. The innocuousness of the inactivated virus product cannot be ensured absolutely. Several recent cases of the disease in Europe have been traced to incompletely inactivated virus. This hazard discourages use of the vaccines in areas where disease is absent, leaving previously unaffected areas devoid of immune protection.
Product instability as well as to lot-to-lot variability stems from the complex and undefined composition of the inactivated virus vaccines. These complex mixtures must be inspected at regular intervals to ensure immunopotency. In some countries this can be as frequent as three times each year. A cold cabin is necessary to maintain the potency of the unstable product, and frequently this is difficult to ensure in some countries.
A potential for deleterious effects is an additional consequence of the complex composition. Anaphylactic shock is an infrequent problem but nevertheless it occurs sufficiently often to pose problems. These disadvantages are reviewed in Brown, Chapter 11 in Molecular Aspects of Picornavirus Infection and Detection, eds Semler and Ehrenfeld, American Society for Microbiology: Washington, DC, 1989, pp 179-191. A further disadvantage stemming from the complex composition of current FMDV vaccines is that it is difficult to distinguish by serological testing animals that have been vaccinated from animals which have been infected (Lubroth et al., Vaccine, 1996; 14: 419-427). A vaccine based on site-specific immunogens would facilitate the design of immunodiagnostic kits to distinguish vaccine-induced antibodies from antibodies induced by infection, and thus ensure the effectiveness of vaccination programs.
The disadvantages of the existing commercial vaccines have encouraged research to replace them with safer and better defined subunit products. However, to be acceptable, such products will need to be of immunogenicity equivalent to that of the currently available inactivated virus vaccines, and will need to provide a wide spectrum of protection against antigenic variants.
Attempts to design a defined subunit vaccine for FMDV began with disruption of the virus particle and identification of its constituent proteins as shown in FIG. 1. This results in a considerable loss of immunogenic activity in comparison to intact inactivated virus particles (Brown, 1989). However, it was found that the separated capsid protein now uniformly designated as VP1 had, by itself, the capacity to elicit neutralizing antibodies and immunity from challenge in cattle and swine, although the activity was subtype specific and many orders of magnitude lower than that of the intact inactivated virus. None of the other virus proteins evoked neutralizing activity (Laporte et al., C R Acad Sci Paris, 1973; 276:3399-3401 and, Bachrach et al., J. Immunol., 1975; 115:1636-1641).
The neutralizing activity of VP1 focused attention on this protein and initially it was seriously considered as a candidate for a genetically engineered subunit vaccine (Kleid et al., Science, 1981; 214:1125-1129). It soon became clear that the immunogenicity of recombinant VP1 was insufficient to protect animals (Brown, 1989; Brown, 1992). Nevertheless, this work led to the discovery of immunogenic determinants on VP1, and the possibility of more potent site-specific subunit vaccines based on synthetic peptides.
VP1 was mapped for immunogenic sites by Strohmaier et al. (3 Gen Virol, 1982; 59:295-306) who tested VP1 cleavage products generated by cyanogen bromide and proteolytic digestion; by Bittle and Lemer (WO 83/03547) who tested synthetic peptides whose amino acid sequences corresponded to regions of highest variability, by Acharya et al. (Nature, 1989, 337:709-711) who identified surface-exposed sites on FMDV by X-ray crystallography, and by Pfaff et al. (EMBO J, 1982; 7:869-874) who selected a peptide that elicits neutralizing antibody by prediction of a prominent amphipathic helical conformation in VP1. All four approaches converged on a loop region of VP1 spanning residues 137-160 as bearing a determinant that elicits neutralizing antibodies.
Serotype-specific peptides corresponding to the 141-160 region of VP1, also termed the "G-H loop", from FMDV isolates belonging to all seven serotypes have now been shown to elicit protective levels of neutralizing antibody in guinea pigs (Francis et al., Immunology, 1990; 69:171-176). Clearly this region contains a dominant immunogenic site which also carries the serotype specificity of the virus. The initial observations of immunogenicity for these 141-160 VP1 peptides was accomplished using synthetic peptides conjugated to carrier protein KLH (keyhole limpet hemocyanin), a procedure which negates the advantage of manufacturing a well-defined synthetic immunogen. However, the development of VP1 synthetic immunogens was advanced by DiMarchi and Brook (U.S. Pat. No. 4,732,971) who showed that two VP1 sequences from a subtype O isolate, joined in a 200-213 Pro-Pro-Ser-141-158-Pro-Cys-Gly chimeric construct, elicited antibody levels which protected cattle against challenge. Nevertheless, the efficacy of this effect was limited by the low immunogenicity of the peptide immunogen. High doses of 1-5 mg were required to achieve protection from homotypic intradermalingual challenge and one of three animals receiving the 5 mg dose was not protected. Practical application demands that a vaccine formulation provide complete protection from both homotypic and heterotypic challenge with smaller amounts of peptide immunogen.
An important element for an effective peptide vaccine for FMDV is the selection of a precise sequence from the G-H loop domain of VP1 that is optimal for presentation of the B cell epitope that is responsible for eliciting neutralizing antibodies. The frame for the optimal presentation of a domain by a peptide can be affected by a shift as small as a single amino acid position (Moore, Chapter 2 in Synthetic Peptides A User's Guide, ed Grant, WH Freeman and Company: New York, 1992, pp 63-67). Preferred synthetic immunogens comprising the immunodominant G-H loop neutralizing determinant have been disclosed including residues 141-160 (WO 83/0357; Francis et al, 1990), 141-158 (U.S. Pat. No. 4,732,971), 142-158 (Parry et al. WO 91/03255), 137-168 (Morgan and Moore, Am J Vet Res, 1990; 51: 40-45), 138-156 (Taboga et al, J Virol, 1997; 71:2602-2614), and any contiguous sequence of 20 amino acids selected from 130-160 (WO 83/03547). Clearly, there is no consensus on a unique segment of VP1 that is optimal for presentation of the G-H loop VP1 neutralizing site.
In addition, the limited variety of B cell epitopes in prior peptide vaccines have not been adequate to stimulate the broadly -reactive neutralizing antibodies needed for protection against the highly variable FMDV. Up to the present time, only antibodies with relatively limited specificities have been elicited and there is a further restriction on the value of even these limited specificities due to the occurrence over time of point mutations giving rise to escape mutants (Taboga et al, 1997). Alternative methods are needed to provide a peptide vaccine with immunogenicity that is sufficiently broad to overcome antigenic variation among multiple strains and temporal variation caused by the emergence of escape mutants.
Another important factor affecting immunogenicity of a synthetic peptide for an FMDV vaccine is presentation to the immune system of a T-helper cell epitope appropriate for the host. Short synthetic peptides can realize their potential as vaccines only if they contain domains that react with helper T-cell receptors and Class II MHC molecules, in addition to antibody binding sites (Babbitt et al., Nature, 1985; 317:359-361). The 141-160 VP1 peptide alone is somewhat immunogenic for guinea pigs and mice, pointing to the presence of T-cell epitopes in the sequence which are appropriate for those species. In contrast to results with those laboratory animals, vaccine trials in cattle and swine suggest that immune responsiveness to the T-cell epitopes of this VP1 domain is too variable for adequate protection among individuals of these economically significant outbred species (Rodriguez et al, Virology, 1994; 205:24-33; and, Taboga et al, J Virol, 1997; 71:2606-2614). Recently, other synthetic peptides corresponding to regions of FMDV structural proteins have been identified which act as T-helper cell sites. A peptide corresponding to VP1 amino acids 21-40 was found to provide T cell help in cattle to the VP1 neutralizing loop sequence (Collen et al., J Immunol, 1991; 146:749-755); however, the presence of this T cell epitope was not sufficient to provide consistent protection in cattle (Taboga et al, 1997). There is still no information regarding the selection of adequate T helper cell determinants nor is the optimal orientation of B- and T-cell epitopes in FMDV peptide vaccines known.
There remains a need for a synthetic peptide-based vaccine for FMDV that is of potency and broadness superior to the classical vaccines. The prior art peptides discussed above do not adequately address this need.
It is an object of the present invention to provide a synthetic peptide FMD vaccine that meets this need. The methods used to attain this goal are:
1. an effective procedure for the identification of high affinity epitopes, PA1 2. the means to represent an enlarged repertoire of B cell and T cell epitopes with a small number of peptide syntheses through combinatorial chemistry, PA1 3. the means to augment the immunogenicity of a B cell target epitope by combining it with a potent T helper cell (Th) epitope and accommodating the Th epitope to the variable immune responsiveness of a population, and PA1 4. the stabilization of desirable conformational features by the introduction of cyclic constraints. PA1 each A is independently an amino acid or a general immunostimulatory sequence; PA1 each B is independently an amino acid or a chemical linker as further defined below; PA1 each Th is independently a sequence of amino acids that constitutes a helper T cell epitope, or an immune enhancing analog or segment thereof; PA1 "FMDV antigen" is a synthetic peptide antigen as defined further below; PA1 X is an amino acid .alpha.-COOH or .alpha.-CONH.sub.2 ; PA1 n is from 0 to about 10; PA1 m is from 1 to about 4; and PA1 o is from 0 to about 10.
Synthetic peptides have been used for "epitope mapping" to identify immunodominant determinants or epitopes on the surface of proteins, for the development of new vaccines, and diagnostics. Epitope mapping employs a series of overlapping peptides corresponding to regions on the protein of interest to identify sites which participate in antibody-immunogenic determinant interaction. Most commonly, epitope mapping employs peptides of relatively short length to precisely detect linear determinants. A fast method of epitope mapping known as "PEPSCAN" is based on the simultaneous synthesis of hundreds of overlapping peptides, of lengths of 8 to 14 amino acids, coupled to solid supports. The coupled peptides are tested for their ability to bind antibodies. This approach is effective in localizing linear determinants, but the immunodominant epitopes needed for vaccine development are usually high affinity discontinuous epitopes and are difficult to define by the PEPSCAN method (Meloen et al., Ann Biol Clin, 1991; 49:231-242).
An alternative method relies on a set of nested and overlapping peptides of multiple lengths ranging from 15 to 60 residues. These longer peptides are synthesized by a laborious series of independent solid-phase peptide syntheses, rather than by the rapid and simultaneous PEPSCAN syntheses. The resulting set of nested and overlapping peptides can then be used in antibody binding studies and experimental immunizations to identify long peptides which best present immunodominant determinants, including simple discontinuous epitopes. This method is exemplified by the studies of Wang for the mapping of immunodominant sites from HTLV I/II (U.S. Pat. No. 5,476,765) and HCV (U.S. Pat. No. 5,106,726); and it was used for the selection of a precise position on the gp 120 sequence for optimum presentation of an HIV neutralizing epitope (Wang et al., Science, 1991; 254:285-288).
Combinatorial peptide libraries have been designed to cover the broad range of sequences embraced by geographical or population differences and temporal variation of hypervariable antigens and to circumvent the limited capacity of individual peptide immunogens. For example, a particular peptide antigen may be designed to contain multiple positional substitutions organized around a structural framework of invariant residues. This combination maintains the conformation of an immunodominant site and accommodates present and anticipated antigenic variation. Such combinatorial peptides, corresponding to hypervariable immunodominant sites of HIV and HCV, have been described as "mixotopes" by Gras-Masse et al. (Peptide Research, 1992; 5:211-216), and as "structured synthetic antigen library" or "SSAL" by Wang et al. (WO 95/11998).
A "structured synthetic antigen library" or SSAL is composed of an ordered set of related peptide sequences embedded within a larger invariant structural framework capable of maintaining the antigenicity, diagnostic value or therapeutic bioactivity of that site.
Equation 1 provides a representation of the SSAL amino acid sequence. EQU AA.sub.1j AA.sub.2j . . . AA.sub.ij . . . AA.sub.xj (Eqn. 1)
The sequence AA.sub.1j to AA.sub.xj represents the amino acid sequence, from the N-terminus to the C-terminus, of the peptides in the SSAL. Each peptide in the library has length x, and x is constant for each library. At each amino acid position i, j varies from 1 to n, where n represents the number of different amino acids known (or desired) at the ith position. The deliberate introduction or natural occurrence of gaps or insertions, e.g. those in the VP1 sequence alignments in Table 1, is accommodated by the synthesis of separate SSALs with different values of x.
Equation 2 is a mathematical representation of the sum of the relative ratios of amino acids at a given variable position i in the SSAL sequence. ##EQU1##
Equation 2 states that at each position i, the sum of the molar fractions of the n individual amino acids is unity. The molar fraction of the jth amino acid at position i, [AAij], is selected by the practitioner, based upon (1) the prevalence of AA.sub.ij among the subtype strains of an FMDV serotype as found in nature, and optionally (2) the desire to include similar amino acids in order to anticipate escape mutants. The amino acids AA.sub.ij will ordinarily be selected from the naturally occurring amino acids, but may include modified amino acids such as ornithine and norleucine and unnatural amino acids such as D-amino acids. By following these formulas, all the variant sequences of an SSAL having length x can be prepared simultaneously in a single synthesis.
The sequence of the SSAL is determined by aligning the primary amino acid sequences of a related family of antigens, and identifying the invariant and variant loci within the alignment. The invariant loci generally represent the structural framework of the SSAL. The degeneracy within the SSAL is determined by the loci within the alignment that harbor different amino acid residues relative to an arbitrary prototype sequence. After determining which amino acids are to be at each invariant position, the degree of degeneracy for the variable positions in the SSAL library is determined from the number of variants that present each alternative amino acid. The SSAL is then synthesized with single amino acids at the invariant positions and with the requisite degeneracies at the variant positions.
The relative ratios of the amino acid residues at the variant positions may be set equal to one another, or they may be represented in unequal proportions. In a preferred embodiment, the amino acid ratios at one or more variant positions are set to reflect the relative prevalence of each amino acid found in the native sequences. The SSAL libraries exemplified in FIGS. 2 and 4 represent such preferred embodiments, wherein the subscripts represent the relative amounts of the individual residues at each variable position in the SSAL. It should be noted that a large fraction of the peptides present in an SSAL may represent combinations of variable amino acids which have not yet been observed in nature. This feature of an SSAL increases the degree to which protection against FMDV infection by new, emerging strains of FMDV is conferred by the peptide compositions of this invention.
Thus, in a simple manner, the specific amino acids, and optionally their frequency of appearance at each position within the SSAL, are defined by the primary sequences of the different antigens that are used in the alignment. In anticipation of escape mutants, amino acids not known to exist in nature at a variant locus within the aligned epitopes can optionally be incorporated into the library. For example, several or all residues of a specific class of amino acids (e.g. hydrophobic, charged, neutral or polar) can be incorporated at any variable locus. The class of amino acids incorporated at a specific locus can be determined by the predominant type of amino acid found at that variant locus within the antigen alignment (WO 95/11998).
Epitopes termed "promiscuous Th" evoke efficient T cell help, and can be combined with B cell epitopes that by themselves are poorly immunogenic to generate potent peptide immunogens. Well-designed promiscuous Th/B cell epitope chimeric peptides are capable of eliciting Th responses and resultant antibody responses in most members of a genetically diverse population expressing diverse MHC haplotypes. Promiscuous Th can be provided by specific sequences derived from potent immunogens of viral and bacterial origin, including measles virus F protein, hepatitis B virus surface antigen, and Chlamydia trachomatis major outer membrane protein. Many known promiscuous Th have been shown to be effective in potentiating a poorly immunogenic peptide corresponding to the decapeptide hormone LHRH (U.S. Pat. No. 5,759,551).
Potent Th epitopes range in size from about 15 to about 50 amino acid residues in length (U.S. Pat. No. 5,759,5 51), and often share common structural features and may contain specific landmark sequences. For example, a common feature is amphipathic helices, which are alpha-helical structures with hydrophobic amino acid residues dominating one face of the helix and with charged and polar residues dominating the surrounding faces (Cease et al., Proc Natl Acad Sci USA, 1987; 84:4249-4253). Th epitopes frequently contain additional primary amino acid patterns such as a Gly or charged residue followed by two to three -hydrophobic residues, followed in turn by a charged or polar residue. This pattern defines what are called Rothbard sequences. Also, Th epitopes often obey the 1, 4, 5, 8 rule, where a positively charged residue is followed by hydrophobic residues at the fourth, fifth and eighth positions after the charged residue. Since all of these structures are composed of common hydrophobic, charged and polar amino acids, each structure can exist simultaneously within a single Th epitope (Partidos et al., 3 Gen Virol, 1991; 72:1293-1299). Most, if not all, of the promiscuous T cell epitopes fit at least one of the periodicities described above. These features may be incorporated into the designs of idealized artificial Th sites, including SSAL Th epitopes that can provide degeneracy at positions associated with recognition of diverse MHC molecules and retain invariant positions. Lists of variable positions and preferred amino acids are available for MHC-binding motifs (Meister et al., Vaccine, 1995; 13:581-591). For example, the degenerate Th epitope described in WO 95/11998 as "SSAL1TH1" was modeled after a promiscuous epitope taken from the F protein of measles virus (Partidos et al., 1991). SSAL1TH1 (WO 95/11998) was designed to be used in tandem with an LHRH target peptide. Like the measles epitope, SSAL1ITH1 follows the Rothbard sequence and the 1, 4, 5, 8 rule.
(See, e.g., positions 6, 9, 10 and 13 in SSAL1TH1):
1 5 10 15 Asp-Leu-Ser-Asp-Leu-Lys-Gly-Leu-Leu-Leu-His-Lys-Leu-Asp-Gly-Leu- Glu Ile Glu Ile Arg Ile Ile Ile Arg Ile Glu Ile Val Val Val Val Val Val Val Phe Phe Phe Phe Phe Phe Phe
Charged residues Glu or Asp are added at position 1 to increase the charge surrounding the hydrophobic face of the Th. The hydrophobic face of the amphipathic helix is then maintained by hydrophobic residues at 2, 5, 8, 9, 10, 13 and 16, with variability at 2, 5, 8, 9, 10, 13 and 16 to provide a facade with the capability of binding to a wide range of MHC restriction elements. The net effect of the SSAL feature is to enlarge the range of immune responsiveness to an artificial Th (WO 95/11998). All variants within such an SSAL are produced simultaneously in a single solid-phase peptide synthesis in tandem with the targeted B cell epitope and other sequences.
Peptide immunogens are generally more flexible than proteins and tend not to retain any preferred structure. Therefore it can be useful to stabilize a peptide immunogen by the introduction of cyclic constraints. A correctly cyclized peptide immunogen can mimic and preserve the conformation of the targeted epitope and thereby evoke antibodies with -reactivities on that site on the authentic molecule (Moore, Chapter 2 in Synthetic Peptides A User's Guide, ed Grant, WH Freeman and Company: New York, 1992, pp 63-67). This has been accomplished for a peptide that represents the FMDV immunodominant VP1 loop domain, by adding cysteine residues at both the amino and carboxy termini and cyclization by oxidation of the two cysteine residues (WO 83/03547). However, there remains a need for more effective peptide analogs of the VP1 loop domain.
Peptides having more immunologically effective conformations are provided by the present invention. These have been obtained by discovering those positions for the cysteine residues that result in a more accurate imitation of the native epitope (Valero et al. Biomedical Peptides, Proteins and Nucleic Acids, 1995; 1:133-140).